Token estimates, API cost, prompt-cache savings, batch discounts, context fit, rate-limit planning and model comparison from your own prices.
payments ON MAINNET (real money) · 8 live endpoints
| Endpoint | Price | What it answers |
|---|---|---|
GET /batch-discount | $0.02 | Compare realtime and batch API cost for a workload with your discount fraction, and the money saved by waiting. |
GET /cache-savings | $0.02 | Prompt-cache savings for a shared prefix: cost with and without caching, break-even requests and percent saved. |
GET /context-fit | $0.02 | Check whether a prompt fits a context window with output reserve, and how many more chat turns fit. |
GET /llm-cost | $0.02 | Cost of LLM API calls from input and output tokens and your per-million-token prices, per request and in total. |
GET /model-compare | $0.02 | Rank 2 to 20 models by cost for a workload from your own prices; flags models whose context window is too small. |
GET /monthly-budget | $0.02 | Forecast monthly and yearly LLM spend from daily requests, tokens and prices, with monthly growth and per-user cost. |
GET /rate-limit-plan | $0.02 | Sustainable requests per minute under RPM and TPM limits, job duration for N requests, and concurrency needed. |
GET /token-estimate | $0.02 | Rough LLM token estimate of text from characters and words, as a low-high range by content type. Not a tokenizer. |
GET /catalog returns this list as JSON. Activity hub · uptime.
Arithmetic on prices, limits and token counts that the CALLER supplies: no provider price list is built in, because prices change. Token estimates are rough heuristics, not a tokenizer; real counts differ by model and language. Not a quote or a billing statement. Every paid response carries a disclaimer field.